BACKGROUND:Fever in travelers returning from tropical areas is a challenging clinical problem due to the broad spectrum of potential underlying causes. We aimed to assess the epidemiology of fever in travelers returning from tropical areas and evaluate the proportion of individuals remaining without a diagnosis. METHODS:This was an observational, retrospective, single-center study conducted between January 2018 and December 2024 including febrile travelers who were tested for malaria with onset of symptoms within 30 days of returning to Italy. Epidemiological, clinical, imaging and laboratory data were collected from health records. A multivariable logistic regression model was built to assess factors associated with remaining without a definite diagnosis. RESULTS:Among the 778 returning travelers included in the study, 446/778 (57%) remained undiagnosed. Diagnostic capacity improved over time, with the proportion of definite diagnoses rising from 29.4% in 2018 to 57.8% in 2024. An arboviral infection was identified in 107 individuals (13.6%), with dengue being the most frequent diagnosis (95 cases, 12.2%), followed by malaria (72 cases, 9.3%) and pneumonia (42 cases, 5.4%). When compared to travelers returning from Africa, travelers from the America and the Western Pacific regions showed lower odds of remaining without a definite diagnosis, respectively aOR 0.45 (95% CI 0.28-0.73, p = 0.001) and aOR 0.49 (95% CI 0.26-0.95, p = 0.034). CONCLUSIONS:Although over half of febrile travelers remained undiagnosed, this proportion declined in recent years, while dengue emerged as the predominant diagnosis in febrile returning travelers.
BACKGROUND:Pericardium is considered electrically inert, but diffuse ST-elevation is an electrocardiographic marker of acute pericarditis. We hypothesised that ST-elevation in acute pericarditis may reflect underlying myocardial involvement. Accordingly, this study aimed to assess the association between ST-elevation and myocardial involvement in pericarditis patients and to further characterise the clinical features and long-term outcomes of myopericarditis compared with isolated pericarditis. METHODS:This longitudinal multicentre study included 351 pericarditis patients (328 recurrent; 180 females), 70/351 with myopericarditis, defined by troponin elevation and/or suggestive cardiac MRI. RESULTS:121 patients had ST-elevation (34.5%); they were younger: 38 years (23-53) vs 47 (31-58) (median (IQR)) (p<0.001), more often male: 63.6% (77/121) vs 40.9% (94/230) (p<0.001) and had higher C reactive protein values: 92.0 (35-170) vs 58.4 mg/L (15.8-137.5) (median (IQR)) (p=0.002) and less frequent pericardial effusions: 71.1% (86/121) vs 83.5% (192/230) (p=0.004).Myocardial involvement was diagnosed in 70/351 (19.9%) patients, occurring more frequently among those with ST-elevation: 26.4% (32/121), compared with those without: 16.5% (38/230) (p=0.035). ST-elevation predicted myocardial involvement with an OR of 1.82 (95% CI 1.07 to 3.10). Compared with isolated pericarditis, patients with myopericarditis were more frequently male: 61.4% (43/70) vs 45.6% (128/281) (p=0.023) and had a higher prevalence of transient systolic dysfunction: 13.5% (7/52) vs 2.1% (3/141) (p=0.004). During follow-up, myopericarditis patients had a lower remission rate: 18.5% (12/65) vs 31.2% (82/263) (p=0.047) and a higher annual hospitalisation rate (median 0.5 vs 0.4/year, p=0.010), while recurrence rates and disease duration were similar. Treatment strategies, including use of corticosteroids and interleukin 1 blockers, were also comparable. CONCLUSIONS:ST-segment elevation in acute pericarditis was associated with myocardial involvement, supporting the concept that the pericardium is electrically inert. Myopericarditis was associated with lower remission rates and slightly higher hospitalisation needs compared to isolated pericarditis, despite otherwise comparable recurrence rates and treatment strategies.
BackgroundBacterial infections remain a major global health burden, causing significant morbidity and mortality. Despite ongoing advances, prompt diagnosis is still hampered by nonspecific host biomarkers and the inherently slow turnaround of traditional microbiological cultures. These limitations often delay the initiation of appropriate treatments. In recent years, affinity-based proteomic approaches have been explored to address this gap. Among them, the Proximity Extension Assay (PEA) has emerged as a promising multiplexed protein quantification tool, capable of simultaneously measuring hundreds of immune and inflammatory proteins with high sensitivity from minimal sample volumes. Such technologies hold the potential to identify novel biomarkers, thereby improving both diagnosis and patient management in bacterial infections.Main bodyIn this systematic scoping review, we examined studies applying PEA-based proteomics to adult bacterial infections. Out of the records screened, ten studies met inclusion criteria. Most were conducted in Europe and North America, relied primarily on plasma samples, and employed commercially available panels enriched for immune and inflammatory mediators. Study quality varied, with some evidence of variability and potential risk of bias. Across the 379 proteins investigated, a subset of proteins were consistently associated with bacterial infections across multiple clinical contexts, whereas others showed limited or no associations.ConclusionsCurrent applications of PEA-based proteomics in adult bacterial infections is limited and largely exploratory. Rather than supporting immediate clinical translation, the available evidence suggests the value of PEA-based approaches for informing biomarker discovery. Future research should prioritize well-designed, longitudinal, and pathogen-stratified studies in clinically relevant settings to strengthen evidence robustness and support the rational development of proteomics-informed diagnostic and translational strategies.
Antimicrobial resistance (AMR) represents an escalating global health threat, demanding diagnostic strategies capable of rapid, accurate, and comprehensive pathogen characterization. Genomic sequencing has transformed our ability to elucidate resistance mechanisms and track their evolution, yet its routine clinical adoption remains limited by cost, workflow constraints, and extended turnaround times. This narrative review examines how artificial intelligence (AI) and machine learning (ML) can enhance and operationalize sequencing-based diagnostics across the clinical microbiology continuum. We summarize current AI applications in whole-genome sequencing for AMR prediction, pan-genome feature extraction, and multicenter model generalizability, including emerging approaches such as federated learning. We then explore AI-driven metagenomic analytics for pathogen detection, resistome profiling, outbreak investigation, and prognostic modeling. Complementary non-genomic technologies, Raman spectroscopy and MALDI-TOF MS, are also evaluated for their potential to deliver rapid resistance profiling when integrated with ML. Finally, we discuss practical barriers, including cost, dataset standardization, interpretability, and regulatory challenges, while outlining future directions toward scalable, explainable, and equitable AI-guided diagnostics. Integrating AI with genomic and rapid phenotypic tools offers a pathway to real-time surveillance, optimized antimicrobial stewardship, and strengthened preparedness against emerging infectious threats.
